The Natural Flow Regime

A paradigm for river conservation and restoration

N. LeRoy Poff, J. David Allan, Mark B. Bain, James R. Karr, Karen L. Prestegaard, Brian D. Richter, Richard E. Sparks, and Julie C. Stromberg

umans have long been fasci- ing. However, current management nated by the dynamism of The ecological integrity approaches often fail to recognize Hfree-flowing waters. Yet we the fundamental scientific principle have expended great effort to tame of river ecosystems that the integrity of flowing water rivers for transportation, water sup- systems depends largely on their natu- ply, flood control, agriculture, and depends on their natural ral dynamic character; as a result, power generation. It is now recog- these methods frequently prevent suc- nized that harnessing of streams and dynamic character cessful river conservation or restora- rivers comes at great cost: Many tion. Streamflow quantity and tim- rivers no longer support socially val- ing are critical components of water ued native species or sustain healthy The extensive ecological degrada- supply, water quality, and the eco- ecosystems that provide important tion and loss of biological diversity logical integrity of river systems. In- goods and services (Naiman et al. resulting from river exploitation is deed, streamflow, which is strongly 1995, NRC 1992). eliciting widespread concern for con- correlated with many critical physi- servation and restoration of healthy cochemical characteristics of rivers, N. LeRoy Poff is an assistant professor river ecosystems among scientists and such as water temperature, channel in the Department of Biology, Colorado the lay public alike (Allan and Flecker geomorphology, and habitat diver- State University, Fort Collins, CO 80523- 1993, Hughes and Noss 1992, Karr sity, can be considered a “master 1878 and formerly senior scientist at Trout Unlimited, Arlington, VA 22209. et al. 1985, TNC 1996, Williams et variable” that limits the distribution J. David Allan is a professor at the School al. 1996). Extirpation of species, clo- and abundance of riverine species of Natural Resources & Environment, sures of fisheries, groundwater deple- (Power et al. 1995, Resh et al. 1988) University of Michigan, Ann Arbor, MI tion, declines in water quality and and regulates the ecological integrity 48109-1115. Mark B. Bain is a research availability, and more frequent and of flowing water systems (Figure 1). scientist and associate professor at the intense flooding are increasingly rec- Until recently, however, the impor- New York Cooperative Fish & Wildlife ognized as consequences of current tance of natural streamflow variabil- Research Unit of the Department of river management and development ity in maintaining healthy aquatic Natural Resources, Cornell University, policies (Abramovitz 1996, Collier ecosystems has been virtually ignored Ithaca, NY 14853-3001. James R. Karr et al. 1996, Naiman et al. 1995). The in a management context. is a professor in the departments of Fish- eries and Zoology, Box 357980, Univer- broad social support in the United Historically, the “protection” of sity of Washington, Seattle, WA 98195- States for the Endangered Species river ecosystems has been limited in 7980. Karen L. Prestegaard is an associate Act, the recognition of the intrinsic scope, emphasizing water quality and professor in the Department of Geology, value of noncommercial native spe- only one aspect of water quantity: University of Maryland, College Park, MD cies, and the proliferation of water- minimum flow. Water resources 20742. Brian D. Richter is national hy- shed councils and riverwatch teams management has also suffered from drologist in the Biohydrology Program, are evidence of society’s interest in the often incongruent perspectives The Nature Conservancy, Hayden, CO maintaining the ecological integrity and fragmented responsibility of 81639. Richard E. Sparks is director of and self-sustaining productivity of agencies (for example, the US Army the River Research Laboratories at the free-flowing river systems. Corps of Engineers and Bureau of Illinois Natural History Survey, Havana, IL 62644. Julie C. Stromberg is an asso- Society’s ability to maintain and Reclamation are responsible for wa- ciate professor in the Department of restore the integrity of river ecosys- ter supply and flood control, the US Plant Biology, Arizona State University, tems requires that conservation and Environmental Protection Agency Tempe, AZ 85281. © 1997 American management actions be firmly and state environmental agencies for Institute of Biological Sciences. grounded in scientific understand- water quality, and the US Fish &

December 1997 769 Figure 1. Flow regime hibit relatively stable hydrographs is of central importance due to high groundwater inputs (Fig- in sustaining the eco- ure 2a), whereas other streams can logical integrity of flow- fluctuate greatly at virtually any time ing water systems. The of year (Figure 2b). In regions with five components of the seasonal precipitation, some streams flow regime—magni- tude, frequency, dura- are dominated by snowmelt, result- tion, timing, and rate ing in pronounced, predictable run- of change—influence off patterns (Figure 2c), and others integrity both directly lack snow accumulation and exhibit and indirectly, through more variable runoff patterns during their effects on other the rainy season, with peaks occur- primary regulators of ring after each substantial storm integrity. Modification event (Figure 2d). of flow thus has cas- Five critical components of the cading effects on the flow regime regulate ecological pro- ecological integrity of rivers. After Karr 1991. cesses in river ecosystems: the mag- nitude, frequency, duration, timing, and rate of change of hydrologic conditions (Poff and Ward 1989, Wildlife Service for water-dependent spective on water management is Richter et al. 1996, Walker et al. species of sporting, commercial, or needed to guide society’s interac- 1995). These components can be used conservation value), making it diffi- tions with rivers. to characterize the entire range of cult, if not impossible, to manage the flows and specific hydrologic phe- entire river ecosystem (Karr 1991). The natural flow regime nomena, such as floods or low flows, However, environmental dynamism that are critical to the integrity of is now recognized as central to sus- The natural flow of a river varies on river ecosystems. Furthermore, by taining and conserving native spe- time scales of hours, days, seasons, defining flow regimes in these terms, cies diversity and ecological integ- years, and longer. Many years of the ecological consequences of par- rity in rivers and other ecosystems observation from a streamflow gauge ticular human activities that modify (Holling and Meffe 1996, Hughes are generally needed to describe the one or more components of the flow 1994, Pickett et al. 1992, Stanford et characteristic pattern of a river’s flow regime can be considered explicitly. al. 1996), and coordinated actions quantity, timing, and variability— are therefore necessary to protect that is, its natural flow regime. Com- • The magnitude of discharge1 at any and restore a river’s natural flow ponents of a natural flow regime can given time interval is simply the variability. be characterized using various time amount of water moving past a fixed In this article, we synthesize exist- series (e.g., Fourier and wavelet) and location per unit time. Magnitude ing scientific knowledge to argue that probability analyses of, for example, can refer either to absolute or to the natural flow regime plays a critical extremely high or low flows, or of relative discharge (e.g., the amount role in sustaining native biodiversity the entire range of flows expressed of water that inundates a floodplain). and ecosystem integrity in rivers. as average daily discharge (Dunne Maximum and minimum magnitudes Decades of observation of the effects and Leopold 1978). In watersheds of flow vary with climate and water- of human alteration of natural flow lacking long-term streamflow data, shed size both within and among regimes have resulted in a well- analyses can be extended statisti- river systems. grounded scientific perspective on cally from gauged streams in the • The frequency of occurrence refers why altering hydrologic variability same geographic area. The frequency to how often a flow above a given in rivers is ecologically harmful (e.g., of large-magnitude floods can be es- magnitude recurs over some speci- Arthington et al. 1991, Castleberry timated by paleohydrologic studies fied time interval. Frequency of oc- et al. 1996, Hill et al. 1991, Johnson of debris left by floods and by studies currence is inversely related to flow et al. 1976, Richter et al. 1997, Sparks of historical damage to living trees magnitude. For example, a 100-year 1995, Stanford et al. 1996, Toth 1995, (Hupp and Osterkamp 1985, Knox flood is equaled or exceeded on aver- Tyus 1990). Current pressing demands 1972). These historical techniques can age once every 100 years (i.e., a on water use and the continuing alter- be used to extend existing hydrologic chance of 0.01 of occurring in any ation of watersheds require scientists records or to provide estimates of given year). The average (median) to help develop management proto- flood flows for ungauged sites. cols that can accommodate economic River flow regimes show regional 1Discharge (also known as streamflow, flow, uses while protecting ecosystem func- patterns that are determined largely or flow rate) is always expressed in dimen- tions. For humans to continue to rely by river size and by geographic varia- sions of volume per time. However, a great on river ecosystems for sustainable tion in climate, geology, topogra- variety of units are used to describe flow, depending on custom and purpose of charac- food production, power production, phy, and vegetative cover. For ex- terization: Flows can be expressed in near- waste assimilation, and flood con- ample, some streams in regions with instantaneous terms (e.g., ft3/s and m3/s) or trol, a new, holistic, ecological per- little seasonality in precipitation ex- over long time intervals (e.g., acre-ft/yr).

770 BioScience Vol. 47 No. 11 flow is determined from a data series of discharges defined over a specific time interval, and it has a frequency of occurrence of 0.5 (a 50% prob- ability). •The duration is the period of time associated with a specific flow condi- tion. Duration can be defined relative to a particular flow event (e.g., a flood- plain may be inundated for a specific number of days by a ten-year flood), or it can be a defined as a composite York: Replace with new expressed over a specified time period Fig. 2 ... supplied (e.g., the number of days in a year when flow exceeds some value). •The timing, or predictability, of flows of defined magnitude refers to the regularity with which they occur. This regularity can be defined for- mally or informally and with refer- ence to different time scales (Poff 1996). For example, annual peak flows may occur with low seasonal predict- ability (Figure 2b) or with high sea- sonal predictability (Figure 2c). •The rate of change, or flashiness, refers to how quickly flow changes from one magnitude to another. At the extremes, “flashy” streams have rapid rates of change (Figure 2b), whereas “stable” streams have slow rates of change (Figure 2a). Figure 2. Flow histories based on long-term, daily mean discharge records. These histories show within- and among-year variation for (a) Augusta Creek, MI, (b) Hydrologic processes and the flow Satilla River, GA, (c) upper Colorado River, CO, and (d) South Fork of the regime. All river flow derives ulti- McKenzie River, OR. Each water year begins on October 1 and ends on September mately from precipitation, but in any 30. Adapted from Poff and Ward 1990. given time and place a river’s flow is derived from some combination of surface water, soil water, and ground- flow patterns. For example, high The natural flow regime organizes water. Climate, geology, topogra- flows due to rainstorms may occur and defines river ecosystems. In riv- phy, soils, and vegetation help to over periods of hours (for permeable ers, the physical structure of the en- determine both the supply of water soils) or even minutes (for imperme- vironment and, thus, of the habitat, and the pathways by which precipi- able soils), whereas snow will melt is defined largely by physical pro- tation reaches the channel. The wa- over a period of days or weeks, which cesses, especially the movement of ter movement pathways depicted in slowly builds the peak snowmelt water and sediment within the chan- Figure 3a illustrate why rivers in flood. As one proceeds downstream nel and between the channel and flood- different settings have different flow within a watershed, river flow reflects plain. To understand the biodiversity, regimes and why flow is variable in the sum of flow generation and rout- production, and sustainability of virtually all rivers. Collectively, over- ing processes operating in multiple river ecosystems, it is necessary to land and shallow subsurface flow small tributary watersheds. The travel appreciate the central organizing role pathways create hydrograph peaks, time of flow down the river system, played by a dynamically varying which are the river’s response to combined with nonsynchronous tribu- physical environment. storm events. By contrast, deeper tary inputs and larger downstream The physical habitat of a river groundwater pathways are respon- channel and floodplain storage ca- includes sediment size and heteroge- sible for baseflow, the form of deliv- pacities, act to attenuate and to neity, channel and floodplain mor- ery during periods of little rainfall. dampen flow peaks. Consequently, phology, and other geomorphic fea- Variability in intensity, timing, annual hydrographs in large streams tures. These features form as the and duration of precipitation (as rain typically show peaks created by wide- available sediment, woody debris, or as snow) and in the effects of spread storms or snowmelt events and other transportable materials are terrain, soil texture, and plant evapo- and broad seasonal influences that moved and deposited by flow. Thus, transpiration on the hydrologic cycle affect many tributaries together habitat conditions associated with combine to create local and regional (Dunne and Leopold 1978). channels and floodplains vary among

December 1997 771 or other features that are left by infrequent high-magnitude floods (e.g., Miller 1990). Over periods of years to decades, a single river can consistently pro- vide ephemeral, seasonal, and per- sistent types of habitat that range from free-flowing, to standing, to no water. This predictable diversity of in-channel and floodplain habitat types has promoted the evolution of species that exploit the habitat mo- saic created and maintained by hy- drologic variability. For many river- ine species, completion of the life cycle requires an array of different habitat types, whose availability over time is regulated by the flow regime (e.g., Greenberg et al. 1996, Reeves et al. 1996, Sparks 1995). Indeed, adaptation to this environmental dy- namism allows aquatic and flood- plain species to persist in the face of seemingly harsh conditions, such as Figure 3. Stream valley cross-sections at various locations in a watershed illustrate basic floods and droughts, that regularly principles about natural pathways of water moving downhill and human influences on destroy and re-create habitat elements. hydrology. Runoff, which occurs when precipitation exceeds losses due to evaporation From an evolutionary perspective, and plant transpiration, can be divided into four components (a): overland flow (1) occurs the pattern of spatial and temporal when precipitation exceeds the infiltration capacity of the soil; shallow subsurface stormflow (2) represents water that infiltrates the soil but is routed relatively quickly to habitat dynamics influences the rela- the stream channel; saturated overland flow (3) occurs where the water table is close to tive success of a species in a particu- the surface, such as adjacent to the stream channel, upstream of first-order tributaries, lar environmental setting. This habi- and in soils saturated by prior precipitation; and groundwater flow (4) represents tat template (Southwood 1977), relatively deep and slow pathways of water movement and provides water to the stream which is dictated largely by flow channel even during periods of little or no precipitation. Collectively, overland and regime, creates both subtle and pro- shallow subsurface flow pathways create the peaks in the hydrograph that are a river’s found differences in the natural his- response to storm events, whereas deeper groundwater pathways are responsible for tories of species in different segments baseflow. Urbanized (b) and agricultural (c) land uses increase surface flow by increasing of their ranges. It also influences the extent of impermeable surfaces, reducing vegetation cover, and installing drainage species distribution and abundance, systems. Relative to the unaltered state, channels often are scoured to greater depth by unnaturally high flood crests and water tables are lowered, causing baseflow to drop. as well as ecosystem function (Poff Side-channels, wetlands, and episodically flooded lowlands comprise the diverse flood- and Allan 1995, Schlosser 1990, plain habitats of unmodified river ecosystems (d). Levees or flood walls (e) constructed Sparks 1992, Stanford et al. 1996). along the banks retain flood waters in the main channel and lead to a loss of floodplain Human alteration of flow regime habitat diversity and function. Dams impede the downstream movement of water and can changes the established pattern of greatly modify a river’s flow regime, depending on whether they are operated for storage natural hydrologic variation and dis- (e) or as “run-of-river,” such as for navigation (f). turbance, thereby altering habitat dynamics and creating new condi- rivers in accordance with both flow 1960). In many streams and rivers tions to which the native biota may characteristics and the type and the with a small range of flood flows, be poorly adapted. availability of transportable materials. bankfull flow can build and main- Within a river, different habitat tain the active floodplain through Human alteration of features are created and maintained stream migration (Leopold et al. flow regimes by a wide range of flows. For ex- 1964). However, the concept of a ample, many channel and floodplain dominant discharge may not be ap- Human modification of natural hy- features, such as river bars and riffle– plicable in all flow regimes (Wolman drologic processes disrupts the dy- pool sequences, are formed and main- and Gerson 1978). Furthermore, in namic equilibrium between the move- tained by dominant, or bankfull, dis- some flow regimes, the flows that ment of water and the movement of charges. These discharges are flows build the channel may differ from sediment that exists in free-flowing that can move significant quantities those that build the floodplain. For rivers (Dunne and Leopold 1978). of bed or bank sediment and that example, in rivers with a wide range This disruption alters both gross- occur frequently enough (e.g., every of flood flows, floodplains may ex- and fine-scale geomorphic features several years) to continually modify hibit major bar deposits, such as that constitute habitat for aquatic the channel (Wolman and Miller berms of boulders along the channel, and riparian species (Table 1). After

772 BioScience Vol. 47 No. 11 Table 1. Physical responses to altered flow regimes.

Source(s) of alteration Hydrologic change(s) Geomorphic response(s) Reference(s) Dam Capture sediment moving Downstream channel erosion and Chien 1985, Petts 1984, 1985, downstream tributary headcutting Williams and Wolman 1984

Bed armoring (coarsening) Chien 1985

Dam, diversion Reduce magnitude and frequency Deposition of fines in gravel Sear 1995, Stevens et al. 1995 of high flows Channel stabilization and Johnson 1994, Williams and narrowing Wolman 1984

Reduced formation of point bars, Chien 1985, Copp 1989, secondary channels, oxbows, Fenner et al. 1985 and changes in channel planform

Urbanization, tiling, drainage Increase magnitude and frequency Bank erosion and channel widening Hammer 1972 of high flows

Downward incision and floodplain Prestegaard 1988 disconnection

Reduced infiltration into soil Reduced baseflows Leopold 1968

Levees and channelization Reduce overbank flows Channel restriction causing Daniels 1960, Prestegaard downcutting et al. 1994

Floodplain deposition and Sparks 1992 erosion prevented

Reduced channel migration and Shankman and Drake 1990 formation of secondary channels

Groundwater pumping Lowered water table levels Streambank erosion and channel Kondolf and Curry 1986 downcutting after loss of vegetation stability such a disruption, it may take centu- tion. More than 85% of the inland many invertebrates and fish, can suf- ries for a new dynamic equilibrium waterways within the continental fer high mortality rates. to be attained by channel and flood- United States are now artificially For many rivers, it is land-use plain adjustments to the new flow controlled (NRC 1992), including activities, including timber harvest, regime (Petts 1985); in some cases, a nearly 1 million km of rivers that are livestock grazing, agriculture, and new equilibrium is never attained, affected by dams (Echeverria et al. urbanization, rather than dams, that and the channel remains in a state of 1989). Dams capture all but the fin- are the primary causes of altered continuous recovery from the most est sediments moving down a river, flow regimes. For example, logging recent flood event (Wolman and with many severe downstream con- and the associated building of roads Gerson 1978). These channel and sequences. For example, sediment- have contributed greatly to degrada- floodplain adjustments are some- depleted water released from dams tion of salmon streams in the Pacific times overlooked because they can can erode finer sediments from the Northwest, mainly through effects be confounded with long-term re- receiving channel. The coarsening of on runoff and sediment delivery sponses of the channel to changing the streambed can, in turn, reduce (NRC 1996). Converting forest or climates (e.g., Knox 1972). Recogni- habitat availability for the many prairie lands to agricultural lands tion of human-caused physical aquatic species living in or using generally decreases soil infiltration changes and associated biological interstitial spaces. In addition, chan- and results in increased overland consequences may require many nels may erode, or downcut, trigger- flow, channel incision, floodplain iso- years, and physical restoration of ing rejuvenation of tributaries, which lation, and headward erosion of the river ecosystem may call for dra- themselves begin eroding and mi- stream channels (Prestegaard 1988). matic action (see box on the Grand grating headward (Chien 1985, Petts Many agricultural areas were drained Canyon flood, page 774). 1984). Fine sediments that are con- by the construction of ditches or tile- Dams, which are the most obvi- tributed by tributaries downstream and-drain systems, with the result ous direct modifiers of river flow, of a dam may be deposited between that many channels have become en- capture both low and high flows for the coarse particles of the streambed trenched (Brookes 1988). flood control, electrical power gen- (e.g., Sear 1995). In the absence of These land-use practices, com- eration, irrigation and municipal high flushing flows, species with life bined with extensive draining of water needs, maintenance of recre- stages that are sensitive to sedimen- wetlands or overgrazing, reduce re- ational reservoir levels, and naviga- tation, such as the eggs and larvae of tention of water in watersheds and,

December 1997 773 impoundments, resulting in great loss of river channel habitat and adjacent A controlled flood in the Grand Canyon floodplain wetlands (Toth 1995). Because levees are designed to pre- ince the Glen Canyon dam first began to store water in 1963, creating vent increases in the width of flow, SLake Powell, some 430 km (270 miles) of the Colorado River, including rivers respond by cutting deeper Grand Canyon National Park, have been virtually bereft of seasonal floods. channels, reaching higher velocities, Before 1963, melting snow in the upper basin produced an average peak or both. discharge exceeding 2400 m3/s; after the dam was constructed, releases Channelization and wetland were generally maintained at less than 500 m3/s. The building of the dam drainage can actually increase the also trapped more than 95% of the sediment moving down the Colorado magnitude of extreme floods, be- River in Lake Powell (Collier et al. 1996). cause reduction in upstream storage This dramatic change in flow regime produced drastic alterations in the capacity results in accelerated water dynamic nature of the historically sediment-laden Colorado River. The delivery downstream. Much of the annual cycle of scour and fill had maintained large sandbars along the river damage caused by the extensive banks, prevented encroachment of vegetation onto these bars, and limited flooding along the Mississippi River bouldery debris deposits from constricting the river at the mouths of in 1993 resulted from levee failure as tributaries (Collier et al. 1997). When flows were reduced, the limited the river reestablished historic con- amount of sand accumulated in the channel rather than in bars farther up nections to the floodplain. Thus, al- the river banks, and shallow low-velocity habitat in eddies used by juvenile though elaborate storage dam and fishes declined. Flow regulation allowed for increased cover of wetland and levee systems can “reclaim” the riparian vegetation, which expanded into sites that were regularly scoured floodplain for agriculture and hu- by floods in the constrained fluvial canyon of the Colorado River; however, man settlement in most years, the much of the woody vegetation that established after the dam’s construction occasional but inevitable large floods is composed of an exotic tree, salt cedar (Tamarix sp.; Stevens et al. 1995). will impose increasingly high disas- Restoration of flood flows clearly would help to steer the aquatic and ter costs to society (Faber 1996). The riparian ecosystem toward its former state and decrease the area of wetland severing of floodplains from rivers and riparian vegetation, but precisely how the system would respond to an also stops the processes of sediment artificial flood could not be predicted. erosion and deposition that regulate In an example of adaptive management (i.e., a planned experiment to the topographic diversity of flood- guide further actions), a controlled, seven-day flood of 1274 m3/s was plains. This diversity is essential for released through the Glen Canyon dam in late March 1996. This flow, maintaining species diversity on roughly 35% of the pre-dam average for a spring flood (and far less than floodplains, where relatively small some large historical floods), was the maximum flow that could pass differences in land elevation result in through the power plant turbines plus four steel drainpipes, and it cost large differences in annual inunda- approximately $2 million in lost hydropower revenues (Collier et al. 1997). tion and soil moisture regimes, which The immediate result was significant beach building: Over 53% of the regulate plant distribution and abun- beaches increased in size, and just 10% decreased in size. Full documenta- dance (Sparks 1992). tion of the effects will continue to be monitored by measuring channel cross-sections and studying riparian vegetation and fish populations. Ecological functions of the natural flow regime instead, route it quickly downstream, and baseflow declines during dry pe- Naturally variable flows create and increasing the size and frequency of riods (Figure 3c). maintain the dynamics of in-channel floods and reducing baseflow levels Whereas dams and diversions af- and floodplain conditions and habi- during dry periods (Figure 3b; Leo- fect rivers of virtually all sizes, and tats that are essential to aquatic and pold 1968). Over time, these prac- land-use impacts are particularly evi- riparian species, as shown schemati- tices degrade in-channel habitat for dent in headwaters, lowland rivers cally in Figure 4. For purposes of aquatic species. They may also iso- are greatly influenced by efforts to illustration, we treat the components late the floodplain from overbank sever channel–floodplain linkages. of a flow regime individually, al- flows, thereby degrading habitat for Flood control projects have short- though in reality they interact in riparian species. Similarly, urban- ened, narrowed, straightened, and complex ways to regulate geomor- ization and suburbanization associ- leveed many river systems and cut phic and ecological processes. In de- ated with human population expan- the main channels off from their flood- scribing the ecological functions as- sion across the landscape create plains (NRC 1992). For example, sociated with the components of a impermeable surfaces that direct channelization of the Kissimmee River flow regime, we pay particular at- water away from subsurface path- above Lake Okeechobee, Florida, by tention to high- and low-flow events, ways to overland flow (and often the US Army Corps of Engineers because they often serve as ecologi- into storm drains). Consequently, transformed a historical 166 km cal “bottlenecks” that present criti- floods increase in frequency and in- meandering river with a 1.5 to 3 km cal stresses and opportunities for a tensity (Beven 1986), banks erode, wide floodplain into a 90 km long wide array of riverine species (Poff and channels widen (Hammer 1972), canal flowing through a series of five and Ward 1989).

774 BioScience Vol. 47 No. 11 The magnitude and frequency of high and low flows regulate numer- ous ecological processes. Frequent, moderately high flows effectively transport sediment through the chan- nel (Leopold et al. 1964). This sedi- ment movement, combined with the force of moving water, exports or- ganic resources, such as detritus and attached algae, rejuvenating the bio- logical community and allowing many species with fast life cycles and good colonizing ability to reestab- lish (Fisher 1983). Consequently, the Figure 4. Geomorphic and ecological functions provided by different levels of flow. composition and relative abundance Water tables that sustain riparian vegetation and that delineate in-channel baseflow habitat are maintained by groundwater inflow and flood recharge (A). Floods of of species that are present in a stream varying size and timing are needed to maintain a diversity of riparian plant species or river often reflect the frequency and aquatic habitat. Small floods occur frequently and transport fine sediments, and intensity of high flows (Meffe maintaining high benthic productivity and creating spawning habitat for fishes (B). and Minckley 1987, Schlosser 1985). Intermediate-size floods inundate low-lying floodplains and deposit entrained sedi- High flows provide further eco- ment, allowing for the establishment of pioneer species (C). These floods also import logical benefits by maintaining eco- accumulated organic material into the channel and help to maintain the characteristic system productivity and diversity. form of the active stream channel. Larger floods that recur on the order of decades For example, high flows remove and inundate the aggraded floodplain terraces, where later successional species establish transport fine sediments that would (D). Rare, large floods can uproot mature riparian trees and deposit them in the channel, otherwise fill the interstitial spaces creating high-quality habitat for many aquatic species (E). in productive gravel habitats (Beschta and Jackson 1979). Floods import and riparian (Nilsen et al. 1984) spe- the successful establishment of non- woody debris into the channel (Keller cies with special behavioral or physi- native species with flow-dependent and Swanson 1979), where it creates ological adaptations that suit them spawning and egg incubation require- new, high-quality habitat (Figure 4; to these harsh conditions. ments, such as striped bass (Morone Moore and Gregory 1988, Wallace The duration of a specific flow saxatilis; Turner and Chadwick and Benke 1984). By connecting the condition often determines its eco- 1972) and brown trout (Salmo trutta; channel to the floodplain, high logical significance. For example, dif- Moyle and Light 1996, Strange et al. overbank flows also maintain ferences in tolerance to prolonged 1992). broader productivity and diversity. flooding in riparian plants (Chapman Seasonal access to floodplain wet- Floodplain wetlands provide impor- et al. 1982) and to prolonged low flow lands is essential for the survival of tant nursery grounds for fish and in aquatic invertebrates (Williams and certain river fishes, and such access export organic matter and organ- Hynes 1977) and fishes (Closs and can directly link high wetland produc- isms back into the main channel (Junk Lake 1996) allow these species to tivity with fish production in the stream et al. 1989, Sparks 1995, Welcomme persist in locations from which they channel (Copp 1989, Welcomme 1992). The scouring of floodplain might otherwise be displaced by 1979). Studies of the effects on stream soils rejuvenates habitat for plant dominant, but less tolerant, species. fishes of both extensive and limited species that germinate only on bar- The timing, or predictability, of floodplain inundation (Finger and ren, wetted surfaces that are free of flow events is critical ecologically Stewart 1987, Ross and Baker 1983) competition (Scott et al. 1996) or because the life cycles of many indicate that some fishes are adapted that require access to shallow water aquatic and riparian species are timed to exploiting floodplain habitats, and tables (Stromberg et al. 1997). Flood- to either avoid or exploit flows of these species decline in abundance resistant, disturbance-adapted ripar- variable magnitudes. For example, when floodplain use is restricted. ian communities are maintained by the natural timing of high or low Models indicate that catch rates and flooding along river corridors, even streamflows provides environmen- biomass of fish are influenced by in river sections that have steep banks tal cues for initiating life cycle tran- both maximum and minimum wet- and lack floodplains (Hupp and sitions in fish, such as spawning land area (Power et al. 1995, Osterkamp 1985). (Montgomery et al. 1983, Nesler et Welcomme and Hagborg 1977), and Flows of low magnitude also pro- al. 1988), egg hatching (Næsje et al. empirical work shows that the area vide ecological benefits. Periods of 1995), rearing (Seegrist and Gard of floodplain water bodies during low flow may present recruitment 1978), movement onto the flood- nonflood periods influences the spe- opportunities for riparian plant spe- plain for feeding or reproduction cies richness of those wetland habi- cies in regions where floodplains are (Junk et al. 1989, Sparks 1995, tats (Halyk and Balon 1983). The frequently inundated (Wharton et Welcomme 1992), or migration up- timing of floodplain inundation is al. 1981). Streams that dry tempo- stream or downstream (Trépanier et important for some fish because mi- rarily, generally in arid regions, have al. 1996). Natural seasonal varia- gratory and reproductive behaviors aquatic (Williams and Hynes 1977) tion in flow conditions can prevent must coincide with access to and avail-

December 1997 775 Table 2. Ecological responses to alterations in components of natural flow regime.a

Flow component Specific alteration Ecological response Reference(s) Magnitude and Increased variation Wash-out and/or stranding Cushman 1985, Petts 1984 frequency Loss of sensitive species Gehrke et al. 1995, Kingsolving and Bain 1993, Travnichek et al. 1995 Increased algal scour and wash-out of Petts 1984 organic matter

Life cycle disruption Scheidegger and Bain 1995

Altered energy flow Valentin et al. 1995 Flow stabilization Invasion or establishment of exotic species, leading to: Local extinction Kupferberg 1996, Meffe 1984 Threat to native commercial species Stanford et al. 1996 Altered communities Busch and Smith 1995, Moyle 1986, Ward and Stanford 1979 Reduced water and nutrients to floodplain plant species, causing: Seedling desiccation Duncan 1993 Ineffective seed dispersal Nilsson 1982 Loss of scoured habitat patches and second- Fenner et al. 1985, Rood et al. ary channels needed for plant establishment 1995, Scott et al. 1997, Shankman and Drake 1990 Encroachment of vegetation into channels Johnson 1994, Nilsson 1982

Timing Loss of seasonal flow peaks Disrupt cues for fish: Spawning Fausch and Bestgen 1997, Montgomery et al. 1993, Nesler et al. 1988 Egg hatching Næsje et al. 1995 Migration Williams 1996 Loss of fish access to wetlands or backwaters Junk et al. 1989, Sparks 1995 Modification of aquatic food web structure Power 1992, Wootton et al. 1996 Reduction or elimination of riparian plant Fenner et al. 1985 recruitment Invasion of exotic riparian species Horton 1977 Reduced plant growth rates Reily and Johnson 1982

Duration Prolonged low flows Concentration of aquatic organisms Cushman 1985, Petts 1984 Reduction or elimination of plant cover Taylor 1982 Diminished plant species diversity Taylor 1982 Desertification of riparian species Busch and Smith 1995, Stromberg composition et al. 1996 Physiological stress leading to reduced plant Kondolf and Curry 1986, Perkins et growth rate, morphological change, al. 1984, Reily and Johnson 1982, or mortality Rood et al. 1995, Stromberg et al. 1992

Prolonged baseflow “spikes” Downstream loss of floating eggs Robertson 1997

Altered inundation duration Altered plant cover types Auble et al. 1994

Prolonged inundation Change in vegetation functional type Bren 1992, Connor et al. 1981 Tree mortality Harms et al. 1980 Loss of riffle habitat for aquatic species Bogan 1993

Rate of change Rapid changes in river stage Wash-out and stranding of aquatic species Cushman 1985, Petts 1984

Accelerated flood recession Failure of seedling establishment Rood et al. 1995 aOnly representative studies are listed here. Additional references are located on the Web at http://lamar.colostate.edu/~poff/natflow.html. ability of floodplain habitats (Wel- ral flow regimes through their “emer- southern floodplain forests (Streng comme 1979). The match of reproduc- gence phenologies”—the seasonal et al. 1989). Productivity of riparian tive period and wetland access also sequence of flowering, seed dispersal, forests is also influenced by flow explains some of the yearly variation germination, and seedling growth. timing and can increase when short- in stream fish community composition The interaction of emergence phe- duration flooding occurs in the grow- (Finger and Stewart 1987). nologies with temporally varying ing season (Mitsch and Rust 1984, Many riparian plants also have environmental stress from flooding Molles et al. 1995). life cycles that are adapted to the or drought helps to maintain high The rate of change, or flashiness, seasonal timing components of natu- species diversity in, for example, in flow conditions can influence spe-

776 BioScience Vol. 47 No. 11 cies persistence and coexistence. In variation in flow regime within and (Ward and Stanford 1979). Many many streams and rivers, particu- among rivers (Figure 2), the same lake fish species have successfully larly in arid areas, flow can change human activity in different locations invaded (or been intentionally estab- dramatically over a period of hours may cause different degrees of change lished in) flow-stabilized river envi- due to heavy storms. Non-native relative to unaltered conditions and, ronments (Moyle 1986, Moyle and fishes generally lack the behavioral therefore, have different ecological Light 1996). Often top predators, adaptations to avoid being displaced consequences. these introduced fish can devastate downstream by sudden floods Flow alteration commonly changes native river fish and threaten com- (Minckley and Deacon 1991). In a the magnitude and frequency of high mercially valuable stocks (Stanford dramatic example of how floods can and low flows, often reducing vari- et al. 1996). In the southwestern benefit native species, Meffe (1984) ability but sometimes enhancing the United States, virtually the entire documented that a native fish, the Gila range. For example, the extreme daily native river fish fauna is listed as topminnow (Poeciliopsis occidentalis), variations below peaking power hy- threatened under the Endangered was locally extirpated by the intro- droelectric dams have no natural Species Act, largely as a consequence duced predatory mosquitofish (Gam- analogue in freshwater systems and of water withdrawal, flow stabiliza- busia affinis) in locations where natu- represent, in an evolutionary sense, tion, and exotic species prolifera- ral flash floods were regulated by an extremely harsh environment of tion. The last remaining strongholds upstream dams, but the native species frequent, unpredictable flow distur- of native river fishes are all in dy- persisted in naturally flashy streams. bance. Many aquatic populations liv- namic, free-flowing rivers, where Rapid flow increases in streams of ing in these environments suffer high exotic fishes are periodically reduced the central and southwestern United mortality from physiological stress, by natural flash floods (Minckley States often serve as spawning cues from wash-out during high flows, and Deacon 1991, Minckley and for native minnow species, whose and from stranding during rapid de- Meffe 1987). rapidly developing eggs are either watering (Cushman 1985, Petts Flow stabilization also reduces the broadcast into the water column or 1984). Especially in shallow shore- magnitude and frequency of overbank attached to submerged structures as line habitats, frequent atmospheric flows, affecting riparian plant species floodwaters recede (Fausch and Best- exposure for even brief periods can and communities. In rivers with con- gen 1997, Robertson in press). More result in massive mortality of bot- strained canyon reaches or multiple gradual, seasonal rates of change in tom-dwelling organisms and subse- shallow channels, loss of high flows flow conditions also regulate the per- quent severe reductions in biological results in increased cover of plant sistence of many aquatic and riparian productivity (Weisberg et al. 1990). species that would otherwise be re- species. Cottonwoods (Populus spp.), Moreover, the rearing and refuge moved by flood scour (Ligon et al. for example, are disturbance species functions of shallow shoreline or 1995, Williams and Wolman 1984). that establish after winter–spring backwater areas, where many small Moreover, due to other related ef- flood flows, during a narrow “win- fish species and the young of large fects of flow regulation, including dow of opportunity” when competi- species are found (Greenberg et al. increased water salinity, non-native tion-free alluvial substrates and wet 1996, Moore and Gregory 1988), vegetation often dominates, such as soils are available for germination. are severely impaired by frequent the salt cedar (Tamarix sp.) in the A certain rate of floodwater reces- flow fluctuations (Bain et al. 1988, semiarid western United States sion is critical to seedling germina- Stanford 1994). In these artificially (Busch and Smith 1995). In alluvial tion because seedling roots must re- fluctuating environments, specialized valleys, the loss of overbank flows main connected to a receding water stream or river species are typically can greatly modify riparian commu- table as they grow downward (Rood replaced by generalist species that nities by causing plant desiccation, and Mahoney 1990). tolerate frequent and large varia- reduced growth, competitive exclu- tions in flow. Furthermore, life cycles sion, ineffective seed dispersal, or of many species are often disrupted failure of seedling establishment Ecological responses to altered and energy flow through the ecosys- (Table 2). flow regimes tem is greatly modified (Table 2). The elimination of flooding may Short-term flow modifications clearly also affect animal species that de- Modification of the natural flow re- lead to a reduction in both the natu- pend on terrestrial habitats. For ex- gime dramatically affects both ral diversity and abundance of many ample, in the flow-stabilized Platte aquatic and riparian species in native fish and invertebrates. River of the United States Great streams and rivers worldwide. Eco- At the opposite hydrologic ex- Plains, the channel has narrowed logical responses to altered flow re- treme, flow stabilization below cer- dramatically (up to 85%) over a gimes in a specific stream or river tain types of dams, such as water period of decades (Johnson 1994). depend on how the components of supply reservoirs, results in artifi- This narrowing has been facilitated flow have changed relative to the cially constant environments that by vegetative colonization of sand- natural flow regime for that particu- lack natural extremes. Although pro- bars that formerly provided nest- lar stream or river (Poff and Ward duction of a few species may in- ing habitat for the threatened pip- 1990) and how specific geomorphic crease greatly, it is usually at the ing plover (Charadius melodius) and ecological processes will respond expense of other native species and and endangered least tern (Sterna to this relative change. As a result of of systemwide species diversity antillarum; Sidle et al. 1992). Sand-

December 1997 777 hill cranes (Grus canadensis), which made the Platte River famous, have abandoned river segments that have narrowed the most (Krapu et al. 1984). Changes in the duration of flow conditions also have significant bio- logical consequences. Riparian plant species respond dramatically to chan- nel dewatering, which occurs fre- quently in arid regions due to surface water diversion and groundwater pumping. These biological and eco- logical responses range from altered leaf morphology to total loss of ri- parian vegetation cover (Table 2). Changes in duration of inundation, independent of changes in annual volume of flow, can alter the abun- dance of plant cover types (Auble et al. 1994). For example, increased duration of inundation has contrib- uted to the conversion of grassland to forest along a regulated Austra- lian river (Bren 1992). For aquatic species, prolonged flows of particu- lar levels can also be damaging. In the regulated Pecos River of New Mexico, artificially prolonged high summer flows for irrigation displace the floating eggs of the threatened Pecos bluntnose shiner (Notropis sinius pecosensis) into unfavorable habitat, where none survive (Robertson in press). Modification of natural flow tim- ing, or predictability, can affect Figure 5. A brief history of flow alteration in the United States. aquatic organisms both directly and indirectly. For example, some native fishes in Norway use seasonal flow ing (Table 2). A shift in timing of threatens most of the native fish fauna peaks as a cue for egg hatching, and peak flows from spring to summer, of the American Southwest (Minckley river regulation that eliminates these as often occurs when reservoirs are and Deacon 1991), and artificially peaks can directly reduce local popu- managed to supply irrigation water, increased rates of change caused by lation sizes of these species (Næsje et has prevented reestablishment of the peaking power hydroelectric dams al. 1995). Furthermore, entire food Fremont cottonwood (Populus on historically less flashy rivers cre- webs, not just single species, may be fremontii), the dominant plant spe- ates numerous ecological problems modified by altered flow timing. In cies in Arizona, because flow peaks (Table 2; Petts 1984). A modified regulated rivers of northern Califor- now occur after, rather than before, rate of change can devastate riparian nia, the seasonal shifting of scouring its germination period (Fenner et al. species, such as cottonwoods, whose flows from winter to summer indi- 1985). Non-native plant species with successful seedling growth depends rectly reduces the growth rate of juve- less specific germination require- on the rate of groundwater recession nile steelhead trout (Oncorhyncus ments may benefit from changes in following floodplain inundation. In mykiss) by increasing the relative flood timing. For example, salt the St. Mary River in Alberta, abundance of predator-resistant in- cedar’s (Tamarix sp.) long seed dis- Canada, for example, rapid draw- vertebrates that divert energy away persal period allows it to establish downs of river stage during spring from the food chain leading to trout after floods occurring any time during have prevented the recruitment of (Wootton et al. 1996). In unregu- the growing season, contributing to its young trees (Rood and Mahoney lated rivers, high winter flows re- abundance on floodplains of the west- 1990). Such effects can be reversed, duce these predator-resistant insects ern United States (Horton 1977). however. Restoration of the spring and favor species that are more pal- Altering the rate of change in flow flood and its natural, slow recession atable to fish. can negatively affect both aquatic in the Truckee River in California Riparian plant species are also and riparian species. As mentioned has allowed the successful establish- strongly affected by altered flow tim- above, loss of natural flashiness ment of a new generation of cotton-

778 BioScience Vol. 47 No. 11 Table 3. Recent projects in which restoration of some component(s) of natural flow regimes has occurred or been proposed for specific ecological benefits.

Location Flow component(s) Ecological purpose(s) Reference Trinity River, CA Mimic timing and magnitude of peak Rejuvenate in-channel gravel habitats; restore Barinaga 1996a flow early riparian succession; provide migration flows for juvenile salmon

Truckee River, CA Mimic timing, magnitude, and duration Restore riparian trees, especially cottonwoods Klotz and Swanson of peak flow, and its rate of change 1997 during recession

Owens River, CA Increase base flows; partially restore Restore riparian vegetation and habitat for Hill and Platts in overbank flows native fishes and non-native brown trout press

Rush Creek, CA (and other Increase minimum flows Restore riparian vegetation and habitat for LADWP 1995 tributaries to Mono Lake) waterfowl and non-native fishes

Oldman River and tributaries, Increase summer flows; reduce rates of Restore riparian vegetation (cottonwoods) Rood et al. 1995 southern Alberta, Canada postflood stage decline; mimic natural and cold-water (trout) fisheries flows in wet years

Green River, UT Mimic timing and duration of peak flow Recovery of endangered fish species; enhance Stanford 1994 and duration and timing of nonpeak other native fishes flows; reduce rapid baseflow fluctu- ations from hydropower generation

San Juan River, UT/NM Mimic magnitude, timing, and duration Recovery of endangered fish species —b of peak flow; restore low winter baseflows

Gunnison River, CO Mimic magnitude, timing, and duration Recovery of endangered fish species —b of peak flow; mimic duration and timing of nonpeak flows

Rio Grande River, NM Mimic timing and duration of flood- Ecosystem processes (e.g., nitrogen flux, Molles et al. 1995 plain inundation microbial activity, litter decomposition)

Pecos River, NM Regulate duration and magnitude of Determine spawning and habitat needs Robertson 1997 summer irrigation releases to mimic for threatened fish species spawning flow “spikes”; maintain minimum flows

Colorado River, AZ Mimic magnitude and timing Restore habitat for endangered fish species Collier et al. 1997 and scour riparian zone

Bill Williams River, AZ Mimic natural flood peak timing Promote establishment of native trees USCOE 1996 (proposed) and duration

Pemigewasset River, NH Reduce frequency (i.e., to no more Enhance native Atlantic salmon recovery FERC 1995 than natural frequency) of high flows during summer low-flow season; reduce rate of change between low and high flows during hydropower cycles

Roanoke River, VA Restore more natural patterning of Increased reproduction of striped bass Rulifson and Manooch monthly flows in spring; reduce rate of 1993 change between low and high flows during hydropower cycles

Kissimmee River, FL Mimic magnitude, duration, rate of Restore floodplain inundation to recover Toth 1995 change, and timing of high- and low- wetland functions; reestablish in-channel flow periods habitats for fish and other aquatic species aJ. Polos, 1997, personal communication. US Fish & Wildlife Service, Arcata, CA. bF. Pfeifer, 1997, personal communication. US Fish & Wildlife Service, Grand Junction, CO. wood trees (Klotz and Swanson primarily on one or a few species opinions of the minimal flow needs 1997). that live in the wetted river channel. for certain fish species (e.g., Larson Most of these methods have the nar- 1981). Recent approaches to row intent of establishing minimum A more sophisticated assessment streamflow management allowable flows. The simplest make of how changes in river flow affect use of easily analyzed flow data, of aquatic habitat is provided by the Methods to estimate environmental assumptions about the regional simi- Instream Flow Incremental Method- flow requirements for rivers focus larity of rivers, and of professional ology (IFIM; Bovee and Milhous

December 1997 779 1978). IFIM combines two models, a overall biological diversity and eco- and native species. Thus, to protect biological one that describes the physi- system function benefit from these pristine or nearly pristine systems, it cal habitat preferences of fishes (and variations in species success (Tilman is necessary to preserve the natural occasionally macroinvertebrates) in et al. 1994). Indeed, experience in hydrologic cycle by safeguarding terms of depth, velocity, and substrate, river restoration clearly shows the against upstream river development and a hydraulic one that estimates impossibility of simultaneously en- and damaging land uses that modify how the availability of habitat for gineering optimal conditions for all runoff and sediment supply in the fish varies with discharge. IFIM has species (Sparks 1992, 1995, Toth watershed. been widely used as an organiza- 1995). A holistic view that attempts Most rivers are highly modified, tional framework for formulating to restore natural variability in eco- of course, and so the greatest chal- and evaluating alternative water logical processes and species success lenges lie in managing and restoring management options related to pro- (and that acknowledges the tremen- rivers that are also used to satisfy duction of one or a few fish species dous uncertainty that is inherent in human needs. Can reestablishing the (Stalnaker et al. 1995). attempting to mechanistically model natural flow regime serve as a useful As a predictive tool for ecological all species in the ecosystem) is neces- management and restoration goal? management, the IFIM modeling sary for ecosystem management and We believe that it can, although to approach has been criticized both in restoration (Franklin 1993). varying degrees, depending on the terms of the statistical validity of its present extent of human interven- physical habitat characterizations Managing toward a natural tion and flow alteration affecting a (Williams 1996) and the limited re- flow regime particular river. Recognizing the alism of its biological assumptions natural variability of river flow and (Castleberry et al. 1996). Field tests The first step toward better incorpo- explicitly incorporating the five com- of its predictions have yielded mixed rating flow regime into the manage- ponents of the natural flow regime results (Morehardt 1986). Although ment of river ecosystems is to recog- (i.e., magnitude, frequency, duration, this approach continues to evolve, nize that extensive human alteration timing, and rate of change) into a both by adding biological realism of river flow has resulted in wide- broader framework for ecosystem (Van Winkle et al. 1993) and by spread geomorphic and ecological management would constitute a expanding the range of habitats changes in these ecosystems. The his- major advance over most present modeled (Stalnaker et al. 1995), in tory of river use is also a history of management, which focuses on mini- practice it is often used only to estab- flow alteration (Figure 5). The early mum flows and on just a few species. lish minimum flows for “important” establishment of the US Army Corps Such recognition would also con- (i.e., game or imperiled) fish species. of Engineers is testimony to the im- tribute to the developing science of But current understanding of river portance that the nation gave to de- stream restoration in heavily altered ecology clearly indicates that fish veloping navigable water routes and watersheds, where, all too often, and other aquatic organisms require to controlling recurrent large floods. physical channel features (e.g., bars habitat features that cannot be main- However, growing understanding of and woody debris) are re-created tained by minimum flows alone (see the ecological impacts of flow alter- without regard to restoring the flow Stalnaker 1990). A range of flows is ation has led to a shift toward an regime that will help to maintain necessary to scour and revitalize appreciation of the merits of free- these re-created features. gravel beds, to import wood and flowing rivers. For example, the Wild Just as rivers have been incremen- organic matter from the floodplain, and Scenic Rivers Act of 1968 recog- tally modified, they can be incre- and to provide access to productive nized that the flow of certain rivers mentally restored, with resulting riparian wetlands (Figure 4). Inter- should be protected as a national improvements to many physical and annual variation in these flow peaks resource, and the recent blossoming biological processes. A list of recent is also critical for maintaining chan- of natural flow restoration projects efforts to restore various components nel and riparian dynamics. For ex- (Table 3) may herald the beginning of a natural flow regime (that is, to ample, imposition of only a fixed of efforts to undo some of the dam- “naturalize” river flow) demon- high-flow level each year would sim- age of past flow alterations. The next strates the scope for success (Table ply result in the equilibration of in- century holds promise as an era for 3). Many of the projects summarized channel and floodplain habitats to renegotiating human relationships in Table 3 represent only partial steps these constant peak flows. with rivers, in which lessons from past toward full flow restoration, but they Moreover, a focus on one or a few experience are used to direct wise and have had demonstrable ecological species and on minimum flows fails informed action in the future. benefits. For example, high flood to recognize that what is “good” for A large body of evidence has flows followed by mimicked natural the ecosystem may not consistently shown that the natural flow regime rates of flow decline in the Oldman benefit individual species, and that of virtually all rivers is inherently River of Alberta, Canada, resulted in what is good for individual species variable, and that this variability is a massive cottonwood recruitment may not be of benefit to the ecosys- critical to ecosystem function and that extended for more than 500 km tem. Long-term studies of naturally native biodiversity. As we have al- downstream from the Oldman Dam. variable systems show that some spe- ready discussed, rivers with highly Dampening of the unnatural flow cies do best in wet years, that other altered and regulated flows lose their fluctuations caused by hydroelectric species do best in dry years, and that ability to support natural processes generation on the Roanoke River in

780 BioScience Vol. 47 No. 11 Virginia has increased juvenile abun- required to make judgments about trols and statistical replication are dances of native striped bass. Mim- specific restoration goals and to work often impossible. Too little attention icking short-duration flow spikes that with appropriate components of the and too few resources have been de- are historically caused by summer natural flow regime to achieve those voted to clarifying how restoring thunderstorms in the regulated Pecos goals. Recognition of the natural flow specific components of the flow re- River of New Mexico has benefited variability and careful identification gime will benefit the entire ecosys- the reproductive success of the Pecos of key processes that are linked to tem. Nevertheless, it is clear that, bluntnose shiner. various components of the flow re- whenever possible, the natural river We also recognize that there are gime are critical to making these system should be allowed to repair scientific limits to how precisely the judgments. and maintain itself. This approach is natural flow regime for a particular Setting specific goals to restore a likely to be the most successful and river can be defined. It is possible to more natural regime in rivers with the least expensive way to restore have only an approximate knowl- altered flows (or, equally important, and maintain the ecological integrity edge of the historic condition of a to preserve unaltered flows in pristine of flow-altered rivers (Stanford et al. river, both because some human ac- rivers) should ideally be a cooperative 1996). Although the most effective tivities may have preceded the instal- process involving river scientists, re- mix of human-aided and natural re- lation of flow gauges, and because source managers, and appropriate covery methods will vary with the climate conditions may have changed stakeholders. The details of this pro- river, we believe that existing knowl- over the past century or more. Fur- cess will vary depending on the spe- edge makes a strong case that restor- thermore, in many rivers, year-to- cific objectives for the river in ques- ing natural flows should be a corner- year differences in the timing and tion, the degree to which its flow stone of our management approach quantity of flow result in substantial regime and other environmental vari- to river ecosystems. variability around any average flow ables (e.g., thermal regime, sediment condition. Accordingly, managing supply) have been altered, and the Acknowledgments for the “average” condition can be social and economic constraints that misguided. For example, in human- are in play. Establishing specific cri- We thank the following people for altered rivers that are managed for teria for flow restoration will be chal- reading and commenting on earlier incremental improvements, restoring lenging because our understanding versions of this paper: Jack Schmidt, a flow pattern that is simply propor- of the interactions of individual flow Lou Toth, Mike Scott, David Wegner, tional to the natural hydrograph in components with geomorphic and Gary Meffe, Mary Power, Kurt years with little runoff may provide ecological processes is incomplete. Fausch, Jack Stanford, Bob Naiman, few if any ecological benefits, be- However, quantitative, river-specific Don Duff, John Epifanio, Lori cause many geomorphic and eco- standards can, in principle, be devel- Robertson, Jeff Baumgartner, Tim logical processes show nonlinear re- oped based on the reconstruction of Randle, David Harpman, Mike sponses to flow. Clearly, half of the the natural flow regime (e.g., Rich- Armbruster, and Thomas Payne. peak discharge will not move half of ter et al. 1997). Restoration actions Members of the Hydropower Re- the sediment, half of a migration- based on such guidelines should be form Coalition also offered construc- motivational flow will not motivate viewed as experiments to be moni- tive comments. Excellent final re- half of the fish, and half of an tored and evaluated—that is, adap- views were provided by Greg Auble, overbank flow will not inundate half tive management—to provide criti- Carter Johnson, an anonymous re- of the floodplain. In such rivers, more cal new knowledge for creative viewer, and the editor of BioScience. ecological benefits would accrue management of natural ecosystem Robin Abell contributed to the de- from capitalizing on the natural be- variability (Table 3). velopment of the timeline in Figure tween-year variability in flow. For To manage rivers from this new 5, and graphics assistance was pro- example, in years with above-aver- perspective, some policy changes are vided by Teresa Peterson (Figure 3), age flow, “surplus” water could be needed. The narrow regulatory fo- Matthew Chew (Figure 4) and Robin used to exceed flow thresholds that cus on minimum flows and single Abell and Jackie Howard (Figure 5). drive critical geomorphic and eco- species impedes enlightened river We also thank the national offices of logical processes. management and restoration, as do Trout Unlimited and American Riv- If full flow restoration is impos- the often conflicting mandates of the ers for encouraging the expression of sible, mimicking certain geomorphic many agencies and organizations that the ideas presented here. We espe- processes may provide some ecologi- are involved in the process. Revi- cially thank the George Gund Foun- cal benefits. Well-timed irrigation sions of laws and regulations, and dation for providing a grant to hold could stimulate recruitment of val- redefinition of societal goals and poli- a one-day workshop, and The Na- ued riparian trees such as cotton- cies, are essential to enable managers ture Conservancy for providing lo- woods (Friedman et al. 1995). Stra- to use the best science to develop ap- gistical support for several of the tegically clearing vegetation from propriate management programs. authors prior to the workshop. river banks could provide new Using science to guide ecosystem sources of gravel for sediment- management requires that basic and starved regulated rivers with reduced applied research address difficult References cited peak flows (e.g., Ligon et al. 1995). questions in complex, real-world set- Abramovitz JN. 1996. Imperiled waters, im- In all situations, managers will be tings, in which experimental con- poverished future: the decline of freshwa-

December 1997 781 ter ecosystems. Washington (DC): World- Experimental flooding in the Grand Can- brown trout (Salmo trutta) and grayling watch Institute. Worldwatch paper nr 128. yon. Scientific American 276: 82–89. (Thymallus thymallus) in the River Vojman, Allan JD, Flecker AS. 1993. Biodiversity con- Connor WH, Gosselink JG, Parrondo RT. 1981. Sweden. Regulated Rivers: Research & servation in running waters. BioScience Comparison of the vegetation of three Loui- Management 12: 287–303. 43: 32–43. siana swamp sites with different flooding Halyk LC, Balon EK. 1983. Structure and eco- Arthington AH, King JM, O’Keefe JH, Bunn regimes. American Journal of Botany 68: logical production of the fish taxocene of a SE, Day JA, Pusey BJ, Bluhdorn DR, Thame 320–331. small floodplain system. Canadian Journal R. 1991. Development of an holistic ap- Copp GH. 1989. The habitat diversity and fish of Zoology 61: 2446–2464. proach for assessing environmental flow reproductive function of floodplain ecosys- Hammer TR. 1972. Stream channel enlarge- requirements of riverine ecosystems. Pages tems. Environmental Biology of Fishes 26: ment due to urbanization. Water Resources 69–76 in Pigram JJ, Hooper BA, eds. Water 1–27. Research 8: 1530–1540. allocation for the environment: proceed- Cushman RM. 1985. Review of ecological ef- Harms WR, Schreuder HT, Hook DD, Brown ings of an international seminar and work- fects of rapidly varying flows downstream CL, Shropshire FW. 1980. The effects of shop. University of New England Armidale from hydroelectric facilities. North Ameri- flooding on the swamp forest in Lake (Australia): The Centre for Water Policy can Journal of Fisheries Management 5: Oklawaha, Florida. Ecology 61: 1412–1421. Research. 330–339. Hill MT, Platts WS. In press. Restoration of Auble GT, Friedman JM, Scott ML. 1994. Daniels RB. 1960. Entrenchment of the willow riparian habitat with a multiple flow regime Relating riparian vegetation to present and drainage ditch, Harrison County, Iowa. in the Owens River Gorge, California. Jour- future streamflows. Ecological Applications American Journal of Science 258: 161–176. nal of Restoration Ecology. 4: 544–554. Duncan RP. 1993. Flood disturbance and the Hill MT, Platts WS, Beschta RL. 1991. Ecologi- Bain MB, Finn JT, Booke HE. 1988. Streamflow coexistence of species in a lowland podocarp cal and geomorphological concepts for regulation and fish community structure. forest, south Westland, New Zealand. Jour- instream and out-of-channel flow require- Ecology 69: 382–392. nal of Ecology 81: 403–416. ments. Rivers 2: 198–210. Barinaga M. 1996. A recipe for river recovery? Dunne T, Leopold LB. 1978. Water in Environ- Holling CS, Meffe GK. 1996. Command and Science 273: 1648–1650. mental Planning. San Francisco: W. H. Free- control and the pathology of natural re- Beschta RL, Jackson WL. 1979. The intrusion man and Co. source management. Conservation Biology of fine sediments into a stable gravel bed. Echeverria JD, Barrow P, Roos-Collins R. 1989. 10: 328–337. Journal of the Fisheries Research Board of Rivers at risk: the concerned citizen’s guide Horton JS. 1977. The development and per- Canada 36: 207–210. to hydropower. Washington (DC): Island petuation of the permanent tamarisk type in Beven KJ. 1986. Hillslope runoff processes and Press. the phreatophyte zone of the Southwest. flood frequency characteristics. Pages 187– Faber S. 1996. On borrowed land: public poli- USDA Forest Service. General Technical 202 in Abrahams AD, ed. Hillslope pro- cies for floodplains. Cambridge (MA): Lin- Report nr RM-43: 124–127. cesses. Boston: Allen and Unwin. coln Institute of Land Policy. Hughes FMR. 1994. Environmental change, Bogan AE. 1993. Freshwater bivalve extinc- Fausch KD, Bestgen KR. 1997. Ecology of disturbance, and regeneration in semi-arid tions (Mollusca: Unionida): a search for fishes indigenous to the central and south- floodplain forests. Pages 321–345 in causes. American Zoologist 33: 599–609. western Great Plains. Pages 131–166 in Millington AC, Pye K, eds. Environmental Bovee KD, Milhous R. 1978. Hydraulic simu- Knopf FL, Samson FB, eds. Ecology and change in drylands: biogeographical and lation in instream flow studies: theory and conservation of Great Plains vertebrates. geomorphological perspectives. New techniques. Ft. Collins (CO): Office of Bio- New York: Springer-Verlag. York: John Wiley & Sons. logical Services, US Fish & Wildlife Service. [FERC] Federal Energy Regulatory Commis- Hughes RM, Noss RF. 1992. Biological diver- Instream Flow Information Paper nr 5, FWS/ sion. 1995. Relicensing the Ayers Island sity and biological integrity: current con- OBS-78/33. hydroelectric project in the Pemigewasset/ cerns for lakes and streams. Fisheries 17: Bren LJ. 1992. Tree invasion of an intermittent Merrimack River Basin. Washington (DC): 11–19. wetland in relation to changes in the flood- Federal Energy Regulatory Commission. Hupp CR, Osterkamp WR. 1985. Bottomland ing frequency of the River Murray, Austra- Final environmental impact statement, vegetation distribution along Passage Creek, lia. Australian Journal of Ecology 17: 395– FERC Project nr 2456–009. Virginia, in relation to fluvial landforms. 408. Fenner P, Brady WW, Patten DR. 1985. Effects Ecology 66: 670–681. Brookes A. 1988. Channelized rivers, perspec- of regulated water flows on regeneration of Johnson WC. 1994. Woodland expansion in tives for environmental management. New Fremont cottonwood. Journal of Range the Platte River, Nebraska: patterns and York: John Wiley & Sons. Management 38: 135–138. causes. Ecological Monographs 64: 45–84. Busch DE, Smith SD. 1995. Mechanisms asso- Finger TR, Stewart EM. 1987. Response of Johnson WC, Burgess RL, Keammerer WR. ciated with decline of woody species in fishes to flooding in lowland hardwood 1976. Forest overstory vegetation and envi- riparian ecosystems of the Southwestern wetlands. Pages 86–92 in Matthews WJ, ronment on the Missouri River floodplain US. Ecological Monographs 65: 347–370. Heins DC, eds. Community and evolution- in North Dakota. Ecological Monographs Castleberry DT, et al. 1996. Uncertainty and ary ecology of North American stream fishes. 46: 59–84. instream flow standards. Fisheries 21: 20– Norman (OK): University of Oklahoma Junk WJ, Bayley PB, Sparks RE. 1989. The 21. Press. flood pulse concept in river-floodplain sys- Chapman RJ, Hinckley TM, Lee LC, Teskey Fisher SG. 1983. Succession in streams. Pages tems. Canadian Special Publication of Fish- RO. 1982. Impact of water level changes on 7–27 in Barnes JR, Minshall GW, eds. Stream eries and Aquatic Sciences 106: 110–127. woody riparian and wetland communities. ecology: application and testing of general Karr JR. 1991. Biological integrity: a long- Vol. 10. Kearneysville (WV): US Fish & ecological theory. New York: Plenum Press. neglected aspect of water resource manage- Wildlife Service. Publication nr OBS-82/83. Franklin JF. 1993. Preserving biodiversity: spe- ment. Ecological Applications 1: 66–84. Chien N. 1985. Changes in river regime after cies, ecosystems, or landscapes? Ecological Karr JR, Toth LA, Dudley DR. 1985. Fish the construction of upstream reservoirs. Applications 3: 202–205. communities of midwestern rivers: a history Earth Surface Processes and Landforms 10: Friedman JM, Scott ML, Lewis WM. 1995. of degradation. BioScience 35: 90–95. 143–159. Restoration of riparian forest using irriga- Keller EA, Swanson FJ. 1979. Effects of large Closs GP, Lake PS. 1996. Drought, differential tion, artificial disturbance, and natural organic material on channel form and flu- mortality and the coexistence of a native seedfall. Environmental Management 19: vial processes. Earth Surface Processes and and an introduced fish species in a south 547–557. Landforms 4: 351–380. east Australian intermittent stream. Envi- Gehrke PC, Brown P, Schiller CB, Moffatt DB, Kingsolving AD, Bain MB. 1993. Fish assem- ronmental Biology of Fishes 47: 17–26. Bruce AM. 1995. River regulation and fish blage recovery along a riverine disturbance Collier M, Webb RH, Schmidt JC. 1996. Dams communities in the Murray-Darling river gradient. Ecological Applications 3: 531– and rivers: primer on the downstream ef- system, Australia. Regulated Rivers: Re- 544. fects of dams. Reston (VA): US Geological search & Management 11: 363–375. Klotz JR, Swanson S. 1997. Managed instream Survey. Circular nr 1126. Greenberg L, Svendsen P, Harby A. 1996. Avail- flows for woody vegetation recruitment, a Collier MP, Webb RH, Andrews ED. 1997. ability of microhabitats and their use by case study. Pages 483–489 in Warwick J,

782 BioScience Vol. 47 No. 11 ed. Symposium proceedings: water resources RJ, Whoriskey FG, Black GA. 1983. Spring lated streams in the United States and an education, training, and practice: opportu- migratory synchrony of salmonid, cato- examination of scale-dependence in some nities for the next century. American Water stomid, and cyprinid fishes in Rivière á la hydrological descriptors. Freshwater Biol- Resources Association, Universities Coun- Truite, Québec. Canadian Journal of Zool- ogy 36: 101–121. cil on Water Resources, American Water ogy 61: 2495–2502. Poff NL, Allan JD. 1995. Functional organiza- Works Association; 29 Jun–3 Jul; Keystone, Moore KMS, Gregory SV. 1988. Response of tion of stream fish assemblages in relation CO. young-of-the-year cutthroat trout to ma- to hydrological variability. Ecology 76: 606– Knox JC. 1972. Valley alluviation in south- nipulations of habitat structure in a small 627. western Wisconsin. Annals of the Associa- stream. Transactions of the American Fish- Poff NL, Ward JV. 1989. Implications of tion of American Geographers 62: 401– eries Society 117: 162–170. streamflow variability and predictability for 410. Morehardt JE. 1986. Instream flow method- lotic community structure: a regional analy- Kondolf GM, Curry RR. 1986. Channel ero- ologies. Palo Alto (CA): Electric Power Re- sis of streamflow patterns. Canadian Jour- sion along the Carmel River, Monterey search Institute. Report nr EPRIEA-4819. nal of Fisheries and Aquatic Sciences 46: County, California. Earth Surface Processes Moyle PB. 1986. Fish introductions into North 1805–1818. and Landforms 11: 307–319. America: patterns and ecological impact. ______. 1990. The physical habitat template of Krapu GL, Facey DE, Fritzell EK, Johnson DH. Pages 27–43 in Mooney HA, Drake JA, eds. lotic systems: recovery in the context of 1984. Habitat use by migrant sandhill cranes Ecology of biological invasions of North historical pattern of spatio-temporal het- in Nebraska. Journal of Wildlife Manage- America and Hawaii. New York: Springer- erogeneity. Environmental Management 14: ment 48: 407–417. Verlag. 629–646. Kupferberg SK. 1996. Hydrologic and geomor- Moyle PB, Light T. 1996. Fish invasions in Power ME. 1992. Hydrologic and trophic con- phic factors affecting conservation of a river- California: do abiotic factors determine trols of seasonal algal blooms in northern breeding frog (Rana boylii). Ecological success? Ecology 77: 1666–1669. California rivers. Archiv für Hydrobiologie Applications 6: 1332–1344. Næsje T, Jonsson B, Skurdal J. 1995. Spring 125: 385–410. Larson HN. 1981. New England flow policy. flood: a primary cue for hatching of river Power ME, Sun A, Parker M, Dietrich WE, Memorandum, interim regional policy for spawning Coregoninae. Canadian Journal Wootton JT. 1995. Hydraulic food-chain New England stream flow recommenda- of Fisheries and Aquatic Sciences 52: 2190– models: an approach to the study of food- tions. Boston: US Fish & Wildlife Service, 2196. web dynamics in large rivers. BioScience 45: Region 5. Naiman RJ, Magnuson JJ, McKnight DM, 159–167. Leopold LB. 1968. Hydrology for urban land Stanford JA. 1995. The freshwater impera- Prestegaard KL. 1988. Morphological controls planning: a guidebook on the hydrologic tive: a research agenda. Washington (DC): on sediment delivery pathways. Pages 533– effects of land use. Reston (VA): US Geo- Island Press. 540 in Walling DE, ed. Sediment budgets. logical Survey. Circular nr 554. [NRC] National Research Council. 1992. Res- Wallingford (UK): IAHS Press. International Leopold LB, Wolman MG, Miller JP. 1964. toration of aquatic systems: science, tech- Association of Hydrological Sciences Publi- Fluvial processes in geomorphology. San nology, and public policy. Washington (DC): cation nr 174. Francisco: W. H. Freeman & Sons. National Academy Press. Prestegaard KL, Matherne AM, Shane B, Ligon FK, Dietrich WE, Trush WJ. 1995. Down- ______. 1996. Upstream: salmon and society in Houghton K, O’Connell M, Katyl N. 1994. stream ecological effects of dams, a geo- the Pacific Northwest. Washington (DC): Spatial variations in the magnitude of the morphic perspective. BioScience 45: 183– National Academy Press. 1993 floods, Raccoon River Basin, Iowa. 192. Nesler TP, Muth RT, Wasowicz AF. 1988. Geomorphology 10: 169–182. [LADWP] Los Angeles Department of Water Evidence for baseline flow spikes as spawn- Reeves GH, Benda LE, Burnett KM, Bisson PA, and Power. 1995. Draft Mono Basin stream ing cues for Colorado Squawfish in the Sedell JR. 1996. A disturbance-based eco- and channel restoration plan. Los Angeles: Yampa River, Colorado. American Fisher- system approach to maintaining and restor- Department of Water and Power. ies Society Symposium 5: 68–79. ing freshwater habitats of evolutionarily Meffe GK. 1984. Effects of abiotic disturbance Nilsen ET, Sharifi MR, Rundel PW. 1984. significant units of anadromous salmonids on coexistence of predator and prey fish Comparative water relations of phreato- in the Pacific Northwest. American Fisher- species. Ecology 65: 1525–1534. phytes in the Sonoran Desert of California. ies Society Symposium 17: 334–349. Meffe GK, Minckley WL. 1987. Persistence Ecology 65: 767–778. Reily PW, Johnson WC. 1982. The effects of and stability of fish and invertebrate assem- Nilsson C. 1982. Effects of stream regulation altered hydrologic regime on tree growth blages in a repeatedly disturbed Sonoran on riparian vegetation. Pages 93–106 in along the Missouri River in North Dakota. Desert stream. American Midland Natural- Lillehammer A, Saltveit SJ, eds. Regulated Canadian Journal of Botany 60: 2410–2423. ist 117: 177–191. rivers. New York: Columbia University Resh VH, Brown AV, Covich AP, Gurtz ME, Li Miller AJ. 1990. Flood hydrology and geomor- Press. HW, Minshall GW, Reice SR, Sheldon AL, phic effectiveness in the central Appala- Perkins DJ, Carlsen BN, Fredstrom M, Miller Wallace JB, Wissmar R. 1988. The role of chians. Earth Surface Processes and Land- RH, Rofer CM, Ruggerone GT, Zim- disturbance in stream ecology. Journal of forms 15: 119–134. merman CS. 1984. The effects of ground- the North American Benthological Society Minckley WL, Deacon JE, ed. 1991. Battle water pumping on natural spring communi- 7: 433–455. against extinction: native fish management ties in Owens Valley. Pages 515–527 in Richter BD, Baumgartner JV, Powell J, Braun in the American West. Tucson (AZ): Uni- Warner RE, Hendrix KM, eds. California DP. 1996. A method for assessing hydro- versity of Arizona Press. riparian systems: ecology, conservation, and logic alteration within ecosystems. Conser- Minckley WL, Meffe GK. 1987. Differential productive management. Berkeley (CA): vation Biology 10: 1163–1174. selection by flooding in stream-fish commu- University of California Press. Richter BD, Baumgartner JV, Wigington R, nities of the arid American Southwest. Pages Petts GE. 1984. Impounded rivers: perspectives Braun DP. 1997. How much water does a 93–104 in Matthews WJ, Heins DC, eds. for ecological management. New York: John river need? Freshwater Biology 37: 231– Community and evolutionary ecology of Wiley & Sons. 249. North American stream fishes. Norman ______. 1985. Time scales for ecological con- Robertson L. In press. Water operations on (OK): University of Oklahoma Press. cern in regulated rivers. Pages 257–266 in the Pecos River, New Mexico and the Mitsch WJ, Rust WG. 1984. Tree growth re- Craig JF, Kemper JB, eds. Regulated streams: Pecos bluntnose shiner, a federally-listed sponses to flooding in a bottomland forest advances in ecology. New York: Plenum minnow. US Conference on Irrigation and in northern Illinois. Forest Science 30: 499– Press. Drainage Symposium. 510. Pickett STA, Parker VT, Fiedler PL. 1992. The Rood SB, Mahoney JM. 1990. Collapse of Molles MC, Crawford CS, Ellis LM. 1995. new paradigm in ecology: implications for riparian poplar forests downstream from Effects of an experimental flood on litter conservation biology above the species level. dams in western prairies: probable causes dynamics in the Middle Rio Grande ripar- Pages 66–88 in Fiedler PL, Jain SK, eds. and prospects for mitigation. Environ- ian ecosystem. Regulated Rivers: Research Conservation biology. New York: Chapman mental Management 14: 451–464. & Management 11: 275–281. & Hall. Rood SB, Mahoney JM, Reid DE, Zilm L. Montgomery WL, McCormick SD, Naiman Poff NL. 1996. A hydrogeography of unregu- 1995. Instream flows and the decline of

December 1997 783 riparian cottonwoods along the St. Mary Stanford JA. 1994. Instream flows to assist the port, review of existing project: Alamo Lake, River, Alberta. Canadian Journal of Botany recovery of endangered fishes of the Upper Arizona. 73: 1250–1260. Colorado River Basin. Washington (DC): Valentin S, Wasson JG, Philippe M. 1995. Ross ST, Baker JA. 1983. The response of fishes US Department of the Interior, National Effects of hydropower peaking on epilithon to periodic spring floods in a southeastern Biological Survey. Biological Report nr 24. and invertebrate community trophic struc- stream. American Midland Naturalist 109: Stanford JA, Ward JV, Liss WJ, Frissell CA, ture. Regulated Rivers: Research & Man- 1–14. Williams RN, Lichatowich JA, Coutant CC. agement 10: 105–119. Rulifson RA, Manooch CS III, eds. 1993. Roa- 1996. A general protocol for restoration of Van Winkle W, Rose KA, Chambers RC. 1993. noke River water flow committee report for regulated rivers. Regulated Rivers: Research Individual-based approach to fish popula- 1991–1993. Albemarle-Pamlico estuarine & Management 12: 391–414. tion dynamics: an overview. Transactions study. Raleigh (NC): US Environmental Pro- Stevens LE, Schmidt JC, Brown BT. 1995. Flow of the American Fisheries Society 122: 397– tection Agency. Project nr APES 93-18. regulation, geomorphology, and Colorado 403. Scheidegger KJ, Bain MB. 1995. Larval fish in River marsh development in the Grand Walker KF, Sheldon F, Puckridge JT. 1995. A natural and regulated rivers: assemblage Canyon, Arizona. Ecological Applications perspective on dryland river ecosystems. composition and microhabitat use. Copeia 5: 1025–1039. Regulated Rivers: Research & Management 1995: 125–135. Strange EM, Moyle PB, Foin TC. 1992. Inter- 11: 85–104. Schlosser IJ. 1985. Flow regime, juvenile abun- actions between stochastic and determinis- Wallace JB, Benke AC. 1984. Quantification of dance, and the assemblage structure of tic processes in stream fish community as- wood habitat in subtropical coastal plains stream fishes. Ecology 66: 1484–1490. sembly. Environmental Biology of Fishes streams. Canadian Journal of Fisheries and ______. 1990. Environmental variation, life 36: 1–15. Aquatic Sciences 41: 1643–1652. history attributes, and community struc- Streng DR, Glitzenstein JS, Harcombe PA. 1989. Ward JV, Stanford JA. 1979. The ecology of ture in stream fishes: implications for envi- Woody seedling dynamics in an East Texas regulated streams. New York: Plenum Press. ronmental management assessment. Envi- floodplain forest. Ecological Monographs Weisberg SB, Janicki AJ, Gerritsen J, Wilson ronmental Management 14: 621–628. 59: 177–204. HT. 1990. Enhancement of benthic Scott ML, Friedman JM, Auble GT. 1996. Stromberg JC, Tress JA, Wilkins SD, Clark S. macroinvertebrates by minimum flow from Fluvial processes and the establishment of 1992. Response of velvet mesquite to a hydroelectric dam. Regulated Rivers: Re- bottomland trees. Geomorphology 14: 327– groundwater decline. Journal of Arid Envi- search & Management 5: 265–277. 339. ronments 23: 45–58. Welcomme RL. 1979. Fisheries ecology of flood- Scott, ML, Auble GT, Friedman JM. 1997. Stromberg JC, Tiller R, Richter B. 1996. Effects plain rivers. New York: Longman. Flood dependency of cottonwood establish- of groundwater decline on riparian vegeta- ______. 1992. River conservation—future pros- ment along the Missouri River, Montana, tion of semiarid regions: the San Pedro pects. Pages 454–462 in Boon PJ, Calow P, USA. Ecological Applications 7: 677–690. River, Arizona, USA. Ecological Applica- Petts GE, eds. River conservation and man- Sear DA. 1995. Morphological and sedimento- tions 6: 113–131. agement. New York: John Wiley & Sons. logical changes in a gravel-bed river follow- Stromberg JC, Fry J, Patten DT. 1997. Marsh Welcomme RL, Hagborg D. 1977. Towards a ing 12 years of flow regulation for hydro- development after large floods in an alluvial, model of a floodplain fish population and power. Regulated Rivers: Research & arid-land river. Wetlands 17: 292–300. its fishery. Environmental Biology of Fishes Management 10: 247–264. Taylor DW. 1982. Eastern Sierra riparian veg- 2: 7–24. Seegrist DW, Gard R. 1972. Effects of floods on etation: ecological effects of stream diver- Wharton CH, Lambou VW, Newsome J, Winger trout in Sagehen Creek, California. Trans- sion. Mono Basin Research Group Contri- PV, Gaddy LL, Mancke R. 1981. The fauna actions of the American Fisheries Society bution nr 6, Report to Inyo National Forest. of bottomland hardwoods in the southeast- 101: 478–482. [TNC] The Nature Conservancy. 1996. Troubled ern United States. Pages 87–160 in Clark Shankman D, Drake DL. 1990. Channel migra- waters: protecting our aquatic heritage. Ar- JR, Benforado J, eds. Wetlands of bottom- tion and regeneration of bald cypress in lington (VA): The Nature Conservancy. land hardwood forests. New York: Elsevier western Tennessee. Physical Geography 11: Tilman D, Downing JA, Wedin DA. 1994. Scientific Publishing Co. 343–352. Does diversity beget stability? Nature 371: Williams JG. 1996. Lost in space: minimum Sidle JG, Carlson DE, Kirsch EM, Dinan JJ. 257–264. confidence intervals for idealized PHABSIM 1992. Flooding mortality and habitat re- Toth LA. 1995. Principles and guidelines for studies. Transactions of the American Fish- newal for least terns and piping plovers. restoration of river/floodplain ecosystems— eries Society 125: 458–465. Colonial Waterbirds 15: 132–136. Kissimmee River, Florida. Pages 49–73 in Williams DD, Hynes HBN. 1977. The ecology Southwood TRE. 1977. Habitat, the templet Cairns J, ed. Rehabilitating damaged eco- of temporary streams. II. General remarks for ecological strategies? Journal of Animal systems. 2nd ed. Boca Raton (FL): Lewis on temporary streams. Internationale Revue Ecology 46: 337–365. Publishers/CRC Press. des gesampten Hydrobiologie 62: 53–61. Sparks RE. 1992. Risks of altering the hydro- Travnichek VH, Bain MB, Maceina MJ. 1995. Williams GP, Wolman MG. 1984. Downstream logic regime of large rivers. Pages 119–152 Recovery of a warmwater fish assemblage effects of dams on alluvial rivers. Reston in Cairns J, Niederlehner BR, Orvos DR, after the initiation of a minimum-flow re- (VA): US Geological Survey. Professional eds. Predicting ecosystem risk. Vol XX. lease downstream from a hydroelectric dam. Paper nr 1286. Advances in modern environmental toxi- Transactions of the American Fisheries So- Williams RN, Calvin LD, Coutant CC, Erho cology. Princeton (NJ): Princeton Scientific ciety 124: 836–844. MW, Lichatowich JA, Liss WJ, McConnaha Publishing Co. Trépanier S, Rodríguez MA, Magnan P. 1996. WE, Mundy PR, Stanford JA, Whitney RR. ______. 1995. Need for ecosystem manage- Spawning migrations in landlocked Atlan- 1996. Return to the river: restoration of ment of large rivers and their floodplains. tic salmon: time series modelling of river salmonid fishes in the Columbia River eco- BioScience 45: 168–182. discharge and water temperature effects. system. Portland (OR): Northwest Power Stalnaker CB. 1990. Minimum flow is a myth. Journal of Fish Biology 48: 925–936. Planning Council. Pages 31–33 in Bain MB, ed. Ecology and Turner JL, Chadwick HK. 1972. Distribution Wolman MG, Gerson R. 1978. Relative scales assessment of warmwater streams: work- and abundance of young-of-the-year striped of time and effectiveness of climate in wa- shop synopsis. Washington (DC): US Fish & bass, Morone saxatilis, in relation to river tershed geomorphology. Earth Surface Pro- Wildlife Service. Biological Report nr 90(5). flow in the Sacramento-San Joaquin estu- cesses and Landforms 3: 189–208. Stalnaker C, Lamb BL, Henriksen J, Bovee K, ary. Transactions of the American Fisheries Wolman MG, Miller JP. 1960. Magnitude and Bartholow J. 1995. The instream flow in- Society 101: 442–452. frequency of forces in geomorphic processes. cremental methodology: a primer for IFIM. Tyus HM. 1990. Effects of altered stream flows Journal of Hydrology 69: 54–74. Ft. Collins (CO): National Biological Ser- on fishery resources. Fisheries 15: 18–20. Wootton JT, Parker MS, Power ME. 1996. vice, US Department of the Interior. Bio- [USCOE] US Army Corps of Engineers, Los Effects of disturbance on river food webs. logical Report nr 29. Angeles District. 1996. Reconnaissance re- Science 273: 1558–1561.

784 BioScience Vol. 47 No. 11